BRYAN GUNER
PROFESSIONAL PORTFOLIO
TABLE OF CONTENTS
Contact Information
Software Skills
Project Experience
& Code
Samples
TECHNICAL
SKILLS &
SOFTWARE
DYNAMIC TIME WARPING TRIGGERED GUITAR EFFEC
TS PLATFORM
• System Architecture:
➢ read in a guitar signal during the ‘learning’ phase and isolate
subsections of a performance needed to generate the DTW
learned-threshold.
➢ then implement an analysis based on a modified
dynamic time
warping (DTW) algorithm, to compare the DTW cost function of
incoming live audio ag
ainst the cost-thresholds determined in
the pre recording phase.
In live performance, guitar
effect pedals
are a versatile
yet limiting
asset, requiring
presence of mind
on the
part
of the performer.
This platform offers an automatic solution to the r
estrictions
that guitar effect pedals pr
esent.
Senior Design Project (T
CNJ)
HOW IT WORKS:
This automation was ac
hieved through the use
of Pure Data, a GUI for audio manipulation
applications, with embedded Python
externals. When the algorithm detects a
match, the platform r
uns the ‘pure’ digitalized
audio signal through custom made PD effects
patches!
Dynamic Time Warping Exter
nal:
•Feed in
two
song performances
for
learning
phase and
obtain Least
Cost Path
(LCP)
for
each sub signal
•Compare Incoming
live
signal
with sub
signal
of one of
the recorded performance
•When LCP
value is less than
or equal
to the
LCP obtained
from
learning
phase, trigger
guitar effect
PURE DATA PATCH DEMOS:
Recording
(.wav) Patch: Reverb Patc
h
Spectral Delay
Patch
(visible part of)Fuzz Patch
Signal sent DI into Scarle
t 2i2 to be
ported into pure data and then back
out the DAC into a Fender Super
Champ X2 FSR 15-watt 1x10" Tube
Combo Amp using the settings below
and was captured by my I phone's mic.
PRESS PLAY
PLATFORM DEMO (WAIT UNTIL 0:18 FOR AUDIO)
This demo uses the reverb patc
h (please
forgive the quality of my guitar performance in this; at the time of
this
recording I hadn’t slept in almost 48 hours in a desperate attempt to meet
multiple concurrent project deadlines!!!)
https://github.com/bgoonz/a-A-Cheat-Sheet.git
BEST PRACTICE
& TOOLS & ANIMATED VSCO
DE
EXTENSION
GUIDE
https://github.com/bgoonz/Javascript-Best-Practices_--Tools
https://github.com/bgoonz/useful-snippets-js.git
GUITAR DELAY PEDAL PROJECT
BRYAN GUNER
DESIGN ELEMENTS
● Analog to Digital Converter
(taking analog guitar input, converting to digital to read
and utilize)
● Rectifier & Pre-Amplifier (to accommodate the input constraints
of the micro controller)
● Digital to Analog Converter
(taking digital signal, converting to analog for output)
● Finite Impulse Response Digital Filter
● 2 Push Buttons
OBJECTIVE
● Design a guitar pedal to take a guitar signal as an input, and output the dela
yed and
echoed sound numerous times
● The pedal must implement a preamplifier in order to manag
e the negative voltage
input
● The pedal must poll f
or updated input data in order to produce a continuous
output
● The pedal must use an ADC in order to read and utilize the input
CURRENT MODEL
Dual Op-Amp (MC1458)
Physical Implementation
of Circuit Design
BLOCK DIAGRAM
Guitar
Code
A/D
Pre
Amp
D/A
x
x
Amplif
ier
PSOC
SCHEMATIC
TIMING DIAGRAM
CURRENT
MODEL CODE
MATLAB SIMULATION
● MATLAB was used to debug and
simulate the gui
tar
delay pedal
● Utilized
FIFO (first
in first out), del
ay and echo
PURE DATA LOOPER
Features:
• tap tempo
• Limiter on both the
input
and the output
to
prevent
intermodulation
distortion
produced
by
superimposed layers.
• “Reverse” function for
fun!
BASIC OPERATION
• The loop is stored in a table in Pd,
which is
played repeatedly.
• While this table is being played, the output
and the current input is also written to it.
• If there is no input the loop copied on itself.
DESIGN
CONSIDERATIONS
• Pure Data processes the audio in blocks of samples
(64 samples by default).
• Any audio event occurring during a block will only
be taken into account at the beginning of the next
block.
• The table’s length must thus be a multiple of the
block size (64 in my patch), so that the end of
the
last block of the loop corresponds exactly to the
beginning of the first bloc
k.
• If the table’s length is not a multiple of the bloc
k
size, audible
pops and clicks will appear while
looping, and gradually turn into a very annoying
noise.
PURE DATA SAMPLER
purpose of this design
assignment
was to implement and simulate both state space
and transfer function descriptions
of a set
of
linearized differential
equations
that model
Modeling and Discrete-Logic Control
of a Quadruple-Tank
System
Bryan Guner
1
and Sean Fernandez
1
; Advisor: Dr.
Ambrose Adegbege
1
1
Department of Electrical
and Computer Engineering, The College of New Jersey,
Ewing, NJ
The aforementioned
quadruple tank
system
consists of
four
water tanks
and two
pumps
that form
a
two-input
two-output system.
The
intention is to
control the
level of the water in the bottom
two tanks while
the
pumps delivers
water to
the top
two tanks
that trickle
down into
the
bottom
two via
two outlets
situated at
the bottom
of the top
two tanks.
The inputs
are
the control
voltages sent
to the
pumps
and the
outputs are
the
sensor
voltages
corresponding
to the water
levels in
the lower
tanks. A
mathematical model
is
developed using
a combination
of the
mass
balance equation,
Bernoulli’s
equation and
the derived
flow
coefficient taking
into account
the
proportionality
splitting
in the pump
system.
The system
is first
modeled
by a
set
of nonlinear
differential
equations which are
then linearized
about an
operating point
that
is comprised
of a water level height argument
and the voltage
to the pumps.
Design/Methods
Conclusion
.
Results
The goal of this design
assignment was
to develop
and
simulate a
mathematical model
of a
quadruple-tanks system
and
subsequently implement discrete-logic control using
the LabVIEW
software package.
The purpose
was
to exercise
our knowledge
of the
differential equations that
represent such
a system,
and to
convert
them into
linearized state-space and
transfer function
descriptions that
could be more easily used
to simulate
said system
in LabVIEW.
Nonlinear Model sub-VI:
LabVIEW Implementation:
Linearized
State-Space Model sub-
VI:
Linearized
Transfer Function Model
sub-VI:
Schematic
of 4 Tank
System:
Linear Equations:Nonlinear
Equations:
State-Space Configuration:
Transfer Function
Configuration:
Quadruple Tank-System
Block Diagram:
Output Water Levels Front Panel:
• Purpose was to implement and simulate both state
space and transfer
function to model behavior
of
system
• Valuable learning experience
in linearization
of
discrete control of systems
• Overall, students
successfully modeled a quadruple-
tank system
utilizing LabVIEW
• Learned about the
use of
subVI’s to reduce
complexity and size of
code
• Close approximation
to working experience
of a
control systems engineer
Introduction
Abstract
XOR GATE USING
CMOS LOGIC
XOR
GATE BACKGROUND
• Either A or B high will output
high
• Both A and B high will output
low
• Both A and B low will output low
PSPICE
CMOS
CIRCUIT
DESIGN
SIMULATION RESULTS (ACTIVE PULL UP
- RISING EDGE)
--
Delay to 50% about
40ns
SIMULATION RESULTS (ACTIVE PULL UP
- FALLING
EDGE)
-Delay until level out
approximately .04us =
40ns
PSPICE RTL CIRCUIT DESIGN
SIMULATION RESULTS (RESISTIVE PULL
UP RISING
EDGE
1K)
SIMULATION RESULTS (RESISTIVE PULL
UP FALLING
EDGE 1K)
SIMULATION RESULTS (RESISTIVE PULL UP RIS
ING
EDGE
100K)
SIMULATION RESULTS (RESISTIVE PULL UP F
ALLING
EDGE 100K)
EXPERIMENTAL
RESULTS (FALLING EDGE)
- Propagation delay between 50%-Input to 50%-Output
approximately 50ns
EXPERIMENTAL
RESULTS (RISING EDGE)
- Propagation delay between 50%-Input to 50%-Output
approximately 50ns
SIMULATION RESULTS (RESISTIVE PULL UP RISING
EDGE
1K)
SIMULATION RESULTS (RESISTIVE PULL UP F
ALLING
EDGE 1K)
SIMULATION RESULTS (RESISTIVE PULL UP RIS
ING
EDGE
100K)
SIMULATION RESULTS (RESISTIVE PULL UP F
ALLING
EDGE 100K)
ONE HANDED KEYBOARD
ENG 142 FUNDAMENTALS OF ENGINEERING
TERM PROJECT ( A
WARDED INNOVATIVE DESIGN)
Ergonomic placement of c
haracters
decided by result of character frequency
counter (written in Java)
as a substitute for the conventional
qwerty layout.
SIGNAL DENOISING USING
WAVELET TRANSFORMATION
BRYAN GUNER
IMAGE DENOISING
CONCEPT
● Compared to other denoising techniques, wavelet a
nd wavelet pac
ket denoising
allows the user to retain features in the data
● Data can be compressed by by setting seemingly unimportant wavelet/wav
elet
packets coefficients to zer
o and reconstructing
data
● Interval-dependent thresholds can be applied to denoise data with non-constant
variance since signal noise is not always uniform in
time
METHODS
● Localize features in your data to different scales
● You can preserve important signal or image
features while removing noise
● The wavelet transform leads to a sparse
representation for many real-world
signals and images
● The wavelet transform concentrates signal and
image features in a few
large-magnitude wavelet coef
ficients
● Wavelet coefficients which are small in value are typically noise and you can
"shrink" those coefficients or remove them without affecting the signal or
image quality.
After you threshold the coef
ficients, you reconstruct the data
using the inverse wavelet transform
MATLAB
IMPLEMENTATION
NOISY SIGNAL SHOWS
ORIGINAL SIGNAL
WITH ADDED WHITE NOISE
MATLAB WAVELET DENOISING METHOD 1
• Symlet Wavelet Family
(Symmetrical)
• Universal Threshold
COEFFICIENT
GRAPH
• Original coefficients
shown in blue
• Denoised coefficients
shown in orange
• Samples(X) Vs. Lev
el(Y)
MATLAB WAVELET
DENOISING
METHOD 2
• Symlet Wavelet
Family
(Symmetrical)
• SURE Method
COEFFICIENT
GRAPH
• Original coefficients
shown in blue
• Denoised coefficients
shown in orange
• Samples(X) Vs. Lev
el(Y)
MATLAB WAVELET
DENOISING
METHOD 3
• Symlet Wavelet
Family
(Symmetrical)
• Minimax Method
COEFFICIENT
GRAPH
• Original coefficients
shown in blue
• Denoised coefficients
shown in orange
• Samples(X) Vs. Lev
el(Y)
